Saturday, February 16, 2013

NGDPLT and Central Bank Discretion

Market Monetarists advocate nominal GDP level targeting as a new monetary regime.   The goal is for nominal GDP to remain on a target growth path, but should nominal GDP fall below or rise above that growth path, the monetary authority would be obligated to return nominal GDP to the target path.

Compare this to flexible inflation targeting.   The central bank generally seeks to have the price level rise a targeted amount from its current level, wherever that happens to be.    If past inflation rates are higher or lower than the targeted amount, that is simply forgotten.   The central bank then tries to get the price level to increase next time at the targeted rate.

However, "flexible" inflation targeting allows the central bank extra discretion.   Perhaps it should not aim at the target inflation rate.   Perhaps it should allow inflation to rise above target for a time.  What exactly is causing inflation to rise?   Is it a temporary change in the supply or demand for some particular product?   Or is it a persistent change in potential output--the productive capacity of the economy?   Is this change due to changes in labor supply?   Or is it due to changes in the productivity of labor?    Perhaps it is a good idea to remember past inflation rates from time to time and offset mistakes.   This time, inflation was below target, so it is better to set it above target now.   How will allowing a temporary increase in inflation impact the central bank's credibility?    Will it cause inflation expectations to become unanchored?   

None of these issues are relevant to a monetary authority constrained to target the growth path of nominal GDP.    As Carney has complained:

The main drawback of an NGDP level target in this regard is that it imposes the arbitrary constraint that prices and real activity must move in equal amounts but opposite directions.

A constraint?   A constraint on whom?   The central banker exercising discretion in manipulating inflation and real output.    Is the constraint arbitrary?    Only in the sense that any nominal anchor is arbitrary.

Flexible inflation targeting gives a central banker discretion to tailor monetary policy precisely to current conditions.   

Consider a decrease in the supply of a good with highly inelastic demand and supply.   The price of the good rises more than in proportion to the decrease in the quantity.    Spending on the good rises.    

As a matter of arithmetic, the price level rises. (ceteris paribus.)    In an inflationary environment, this is a shift to a higher growth path for the price level and a transitory increase in inflation.   

Similarly, as a matter of arithmetic,  potential output falls.(ceteris paribus.)   In a growing economy, this is a shift to a lower growth path of potential output, and implies a transitory slowdown in the growth rate of potential output.   In the extreme, this could involve negative growth, a temporary decrease in potential output.

The increase in the price of the good results in an increase in quantity supplied.    This partial offset of the initial supply shock implies a shift of resources away from producing other goods and services.   With supply inelastic, this is a small effect.   

The rest of the story depends entirely on the monetary regime.    With price level targeting, the central bank must force other prices to grow more slowly so that the price level remains on its target growth path.    It does this by slowing spending for  the entire economy.    The demand for the good with the decrease in supply sees slower growth in demand, like every other product, but for the most part, the impact is slower demand growth and smaller price increases in the rest of the economy.   The slower growth in spending will likely result in slower growth in output and employment for the rest of the economy as well.     This has the beneficial effect of freeing up resources to allow for the expansion of the quantity supplied for the good with reduced supply.    Because supply is inelastic, this is a small effect.

Nominal incomes, including wages, will also need to grow more slowly in in aggregate.   Because its demand is inelastic by assumption, nominal incomes increase for those producing the good with reduced supply, so the decrease in nominal incomes must be concentrated in the rest of the economy.  Fortunately, lower nominal wages and other incomes will allow production to recover in the rest of the economy, at least in the long run.

With inflation targeting, if this event has occurred like a bolt from the blue, the central bank ignores the increase in the inflation rate and upward shift in the growth path of prices.    It seeks to have prices rise at the targeted growth rate in the future.   There is no need for a slow down price increases in the rest of the economy.   There is no need to contract nominal GDP growth.   

For example, if a central bank is surprised by a bad corn harvest, the observed increase in inflation is ignored.  The growth path of nominal GDP rises by the increase in spending on corn.  (The increase in spending on corn follows from the assumption of inelastic demand.   If spending in the rest of the economy remains on a constant growth path, nominal GDP must shift up to a higher growth path.)

However, if the decrease in supply is foreseen, then strict inflation targeting requires that the central bank act as it would with a price level target.   It must slow the growth of spending on output so that the expected rapid increase in the price of the good with decreased supply is offset by slower increases in the prices of other goods and services.   Nominal incomes and nominal wages must grow more slowly.  

For example, if rapid economic growth in China is putting persistent upward pressure on the price of crude oil, the Fed would need to restrain nominal GDP growth in the U.S. so that the prices of other goods and services grow more slowly.   Wages and other nominal incomes would need to grow more slowly reflecting the slower growth in real incomes in the U.S. due to increased competition for crude oil by developing countries.

With flexible inflation targeting, the central bank can consider the demand and supply elasticities of the good expected to face the supply shock, and determine what is the optimal increase in the price level.    The arithmetic increase in the price level or growth path of prices, and transitory inflation could be allowed, with real output only falling by the small decrease in potential output.  

With a nominal GDP level target, the response is determined by the rule, and it is clearly less than optimal.  With spending on the good that suffered the decrease in supply rising, spending in the rest of the economy must fall.   With growing nominal GDP, this implies that spending on the good that had the decrease in supply must shift to a higher growth path, while spending in the rest of the economy must shift to a lower growth path.  

This will tend to have an effect similar to a price level target, except that with sticky prices, the slowdown in spending in the rest of the economy is less than what would be necessary to keep inflation on target.    (The slowdown in production also "counts" as reduced nominal GDP. )   Further, nominal income as a whole does not decrease, but the higher nominal incomes earned in the sector with reduced supply must be offset by slower growth of nominal income in the rest of the economy.   Reduced spending, production, and employment in the rest of the economy has the beneficial consequence of freeing up resources to allow an increase in quantity supplied of the good with reduced supply.   Unfortunately, because supply is inelastic by assumption, that is a small benefit.

What would be necessary for a nominal GDP level target to be optimal?    It is instructive to consider one situation where the rule is optimal.   If the demand for the good with the decreased supply was unit elastic, and the supply was perfectly inelastic, then a nominal GDP target would be perfect.   Spending on the good with the decreased supply is constant, or rather, on a stable growth path.   This allows spending in the rest of the economy to remain on a stable growth path as well.   Because supply is perfectly inelastic, there is no increase in quantity supplied, and so no need to shift resources away from the rest of the economy to partly offset the decrease in supply.    There is no need to signal firms producing other products that they should slightly curtail production to free up resources.   

The rest of the economy would be sheltered from the consequences of the supply shock.   Of course, real incomes are lower.   The reduced supply, higher price, and reduced quantity of the one product clearly makes everyone worse off.   What it is avoided is disruption to other parts of the economy.

Keeping the assumption that supply is perfectly inelastic, if demand for the good with reduced supply  was  inelastic, the central bank should allow nominal GDP to shift to a higher growth path.   If, on the other hand, the demand the good is elastic, then the central bank needs a contractionary policy, reducing the growth path of nominal GDP.    At least if the central bank's goal is to avoid disruption to the rest of the economy.

If supply is not perfectly inelastic, then nominal GDP level targeting is only optimal if the demand for a good with reduced supply is at least slightly inelastic.   The less inelastic (more elastic) is supply, the more inelastic would demand need to be.  This provides the proper signals and incentives to shift resources away from other sectors of the economy and expand the quantity supplied of the good with the decrease in supply.  

We can at least imagine a central bank adjusting the growth rate of nominal GDP in a way that manipulates spending in the rest of the economy such that the appropriate amount of resources are freed up to provide the needed adjustment in quantity supplied.   This implies slight changes in inflation in the rest of the economy, which when combined with the inflation generated by the supply shock, suggests a varying target for inflation.   Doesn't it follow that a central bank with an optimal monetary policy must have a flexible inflation target?

Why then, do I still favor nominal GDP level targeting?   Because "flexible" inflation targeting, where central bankers anticipate which goods will face changes in supply, and then carefully consider the supply and demand elasticities, and so allow nominal GDP and inflation to adjust the optimal amount, is entirely unrealistic.   No one has anything like the information needed to accomplish this.   This is the old problem of economists imagining a perfectly omniscient and benevolent despot creating the perfect regulation to correct market failures.   Nothing like that exists.

What is the best, which I consider the same as, least bad, macroeconomic environment for microeconomic coordination?     Maximizing social welfare is not a rule for monetary policy.   What does flexible inflation targeting mean?    Target inflation when that maximizes social welfare, and then allow inflation to deviate from target when that results in more social welfare.   That isn't inflation targeting of any sort, that is giving central bankers discretion to do whatever they think is best.

I am not surprised that central bankers have a natural tendency to favor a policy that simply allows them to do whatever it is that they think is best.   It is natural for them to dislke "arbitrary constraints."   During the Great Moderation, we could pretend that discretion would work just fine.  In my view, the Great Recession proves that we need something new.   Flexible inflation targeting is a proven failure.

HT Scott Sumner

No comments:

Post a Comment